Showing 1–10 of 10 results for author: Raghav, S

AfroXLMR-Comet: Multilingual Knowledge Distillation with Attention Matching for Low-Resource languages

Authors: Joshua Sakthivel Raju, Sanjay S, Jaskaran Singh Walia, Srinivas Raghav, Vukosi Marivate

Abstract: Language model compression through knowledge distillation has emerged as a promising approach for deploying large language models in resource-constrained environments. However, existing methods often struggle to maintain performance when distilling multilingual models, especially for low-resource languages. In this paper, we present a novel hybrid distillation approach that combines traditional kn… ▽ More Language model compression through knowledge distillation has emerged as a promising approach for deploying large language models in resource-constrained environments. However, existing methods often struggle to maintain performance when distilling multilingual models, especially for low-resource languages. In this paper, we present a novel hybrid distillation approach that combines traditional knowledge distillation with a simplified attention matching mechanism, specifically designed for multilingual contexts. Our method introduces an extremely compact student model architecture, significantly smaller than conventional multilingual models. We evaluate our approach on five African languages: Kinyarwanda, Swahili, Hausa, Igbo, and Yoruba. The distilled student model; AfroXLMR-Comet successfully captures both the output distribution and internal attention patterns of a larger teacher model (AfroXLMR-Large) while reducing the model size by over 85%. Experimental results demonstrate that our hybrid approach achieves competitive performance compared to the teacher model, maintaining an accuracy within 85% of the original model's performance while requiring substantially fewer computational resources. Our work provides a practical framework for deploying efficient multilingual models in resource-constrained environments, particularly benefiting applications involving African languages. △ Less

Submitted 25 February, 2025; originally announced February 2025.

Comparative Analysis of Black Hole Mass Estimation in Type-2 AGNs: Classical vs. Quantum Machine Learning and Deep Learning Approaches

Authors: Sathwik Narkedimilli, Venkata Sriram Amballa, N V Saran Kumar, R Arun Kumar, R Praneeth Reddy, Satvik Raghav, Manish M, Aswath Babu H

Abstract: In the case of Type-2 AGNs, estimating the mass of the black hole is challenging. Understanding how galaxies form and evolve requires considerable insight into the mass of black holes. This work compared different classical and quantum machine learning (QML) algorithms for black hole mass estimation, wherein the classical algorithms are Linear Regression, XGBoost Regression, Random Forest Regresso… ▽ More In the case of Type-2 AGNs, estimating the mass of the black hole is challenging. Understanding how galaxies form and evolve requires considerable insight into the mass of black holes. This work compared different classical and quantum machine learning (QML) algorithms for black hole mass estimation, wherein the classical algorithms are Linear Regression, XGBoost Regression, Random Forest Regressor, Support Vector Regressor (SVR), Lasso Regression, Ridge Regression, Elastic Net Regression, Bayesian Regression, Decision Tree Regressor, Gradient Booster Regressor, Classical Neural Networks, Gated Recurrent Unit (GRU), LSTM, Deep Residual Networks (ResNets) and Transformer-Based Regression. On the other hand, quantum algorithms including Hybrid Quantum Neural Networks (QNN), Quantum Long Short-Term Memory (Q-LSTM), Sampler-QNN, Estimator-QNN, Variational Quantum Regressor (VQR), Quantum Linear Regression(Q-LR), QML with JAX optimization were also tested. The results revealed that classical algorithms gave better R^2, MAE, MSE, and RMSE results than the quantum models. Among the classical models, LSTM has the best result with an accuracy of 99.77%. Estimator-QNN has the highest accuracy for quantum algorithms with an MSE of 0.0124 and an accuracy of 99.75%. This study ascertains both the strengths and weaknesses of the classical and the quantum approaches. As far as our knowledge goes, this work could pave the way for the future application of quantum algorithms in astrophysical data analysis. △ Less

Submitted 24 February, 2025; v1 submitted 21 February, 2025; originally announced February 2025.

Comments: 29 pages, 12 Figures, 6 Tables

FL-DABE-BC: A Privacy-Enhanced, Decentralized Authentication, and Secure Communication for Federated Learning Framework with Decentralized Attribute-Based Encryption and Blockchain for IoT Scenarios

Authors: Sathwik Narkedimilli, Amballa Venkata Sriram, Satvik Raghav

Abstract: This study proposes an advanced Federated Learning (FL) framework designed to enhance data privacy and security in IoT environments by integrating Decentralized Attribute-Based Encryption (DABE), Homomorphic Encryption (HE), Secure Multi-Party Computation (SMPC), and Blockchain technology. Unlike traditional FL, our framework enables secure, decentralized authentication and encryption directly on… ▽ More This study proposes an advanced Federated Learning (FL) framework designed to enhance data privacy and security in IoT environments by integrating Decentralized Attribute-Based Encryption (DABE), Homomorphic Encryption (HE), Secure Multi-Party Computation (SMPC), and Blockchain technology. Unlike traditional FL, our framework enables secure, decentralized authentication and encryption directly on IoT devices using DABE, allowing sensitive data to remain locally encrypted. Homomorphic Encryption permits computations on encrypted data, and SMPC ensures privacy in collaborative computations, while Blockchain technology provides transparent, immutable record-keeping for all transactions and model updates. Local model weights are encrypted and transmitted to fog layers for aggregation using HE and SMPC, then iteratively refined by the central server using differential privacy to safeguard against data leakage. This secure, privacy-preserving FL framework delivers a robust solution for efficient model training and real-time analytics across distributed IoT devices, offering significant advancements in secure decentralized learning for IoT applications. △ Less

Submitted 26 October, 2024; originally announced October 2024.

Photometric Analysis for Predicting Star Formation Rates in Large Galaxies Using Machine Learning and Deep Learning Techniques

Authors: Satvik Raghav, Prasanth Ayitapu, Sathwik Narkedimilli, Sujith Makam, Aswath Babu H

Abstract: Star formation rates (SFRs) are a crucial observational tracer of galaxy formation and evolution. Spectroscopy, which is expensive, is traditionally used to estimate SFRs. This study tests the possibility of inferring SFRs of large samples of galaxies from only photometric data, using state-of-the-art machine learning and deep learning algorithms. The dataset adopted in this work is the one collec… ▽ More Star formation rates (SFRs) are a crucial observational tracer of galaxy formation and evolution. Spectroscopy, which is expensive, is traditionally used to estimate SFRs. This study tests the possibility of inferring SFRs of large samples of galaxies from only photometric data, using state-of-the-art machine learning and deep learning algorithms. The dataset adopted in this work is the one collected by Delli Veneri et al. (2019): it includes photometric data of more than 27 million galaxies coming from the Sloan Digital Sky Survey Data Release 7 (SDSS-DR7). The algorithms we implemented and tested for comparing the performances include Linear Regression, Long Short-Term Memory (LSTM) networks, Support Vector Regression (SVR), Random Forest Regressor, Decision Tree Regressor, Gradient Boosting Regressor, and classical deep learning models. Our results mention that the Linear Regression model predicted an impressive accuracy of 98.97 percent as measured by the Mean Absolute Error (MAE), demonstrating that machine-learning approaches can be effective when it comes to photometric SFR estimation. Besides, the paper also reported the results for other intelligent algorithms, which predicted the SFRs, providing a detailed comparison of the performance of different machine learning algorithms in the photometric SFR estimation. This study not only shows the estimated SFR from photometric data is promising but also opens a door toward the application of machine learning and deep learning in astrophysics. △ Less

Submitted 9 October, 2024; originally announced October 2024.

Comments: 19 pages, 6 figures, and 2 tables

Predicting Stellar Metallicity: A Comparative Analysis of Regression Models for Solar Twin Stars

Authors: Sathwik Narkedimilli, Satvik Raghav, Sujith Makam, Prasanth Ayitapu, Aswath Babu H

Abstract: The research focuses on determining the metallicity ([Fe/H]) predicted in the solar twin stars by using various regression modeling techniques which are, Random Forest, Linear Regression, Decision Tree, Support Vector, and Gradient Boosting. The data set that is taken into account here includes Stellar parameters and chemical abundances derived from a high-accuracy abundance catalog of solar twins… ▽ More The research focuses on determining the metallicity ([Fe/H]) predicted in the solar twin stars by using various regression modeling techniques which are, Random Forest, Linear Regression, Decision Tree, Support Vector, and Gradient Boosting. The data set that is taken into account here includes Stellar parameters and chemical abundances derived from a high-accuracy abundance catalog of solar twins from the GALAH survey. To overcome the missing values, intensive preprocessing techniques involving, imputation are done. Each model will subjected to training using different critical observables, which include, Mean Squared Error(MSE), Mean Absolute Error(MAE), Root Mean Squared Error(RMSE), and R-squared. Modeling is done by using, different feature sets like temperature: effective temperature(Teff), surface gravity: log g of 14-chemical-abundances namely, (([Na/Fe], [Mg/Fe], [Al/Fe], [Si/Fe], [Ca/Fe], [Sc/Fe], [Ti/Fe], [Cr/Fe], [Mn/Fe], [Ni/Fe], [Cu/Fe], [Zn/Fe], [Y/Fe], [Ba/Fe])). The target variable considered is the metallicity ([Fe/H]). The findings indicate that the Random Forest model achieved the highest accuracy, with an MSE of 0.001628 and an R-squared value of 0.9266. The results highlight the efficacy of ensemble methods in handling complex datasets with high dimensionality. Additionally, this study underscores the importance of selecting appropriate regression models for astronomical data analysis, providing a foundation for future research in predicting stellar properties with machine learning techniques. △ Less

Submitted 9 October, 2024; originally announced October 2024.

Comments: 16 pages, 7 figure, and 2 tables

Dispersion Managed Elliptical Atomtronics for Interferometry

Authors: Sriganapathy Raghav, Suranjana Ghosh, Jayanta Bera, Utpal Roy

Abstract: Circular atomtronics is known to exhibit a uniform ground state, unlike elliptical atomtronics. In elliptical atomtronics, the matter wave tends to accumulate along the semimajor edges during its time dynamics, which we depict by the survival function. Consequently, the dynamical time scales become coupled to the eccentricity, making the dynamics nontrivial for applications. We report that an appr… ▽ More Circular atomtronics is known to exhibit a uniform ground state, unlike elliptical atomtronics. In elliptical atomtronics, the matter wave tends to accumulate along the semimajor edges during its time dynamics, which we depict by the survival function. Consequently, the dynamical time scales become coupled to the eccentricity, making the dynamics nontrivial for applications. We report that an appropriate dispersion management can decouple the time scales from the eccentricity. One can choose the suitable dispersion coefficient from the overlap function involving the corresponding ground state. We focus on producing distinct fractional matter waves inside an elliptical waveguide to achieve efficient atom interferometry. The said dispersion engineering can recover fractional revivals in the elliptical waveguide, analogous to the circular case. We demonstrate atom interferometry for the engineered elliptical atomtronics, where matter wave interference is mediated by an external harmonic trap for controlled interference patterns. △ Less

Submitted 2 December, 2024; v1 submitted 13 April, 2024; originally announced April 2024.

Comments: 11 pages, 7 figures

Journal ref: Eur. Phys. J. Plus 139, 1047 (2024)

Nonlinearity mediated miscibility dynamics of mass-imbalanced binary Bose Einstein condensate for circular atomtronics

Authors: Sriganapathy Raghav, Suranjana Ghosh, Barun Halder, Utpal Roy

Abstract: We explore the nonlinearity-induced and fractional revivals-driven miscibility dynamics of quasi-2D mass-imbalanced binary Bose-Einstein condensates, confined in a ring-shaped waveguide. During their time-evolution, the two condensate species generally remain miscible, as observed in the spatial density distributions and the autocorrelation functions. Although, the investigation is carried out for… ▽ More We explore the nonlinearity-induced and fractional revivals-driven miscibility dynamics of quasi-2D mass-imbalanced binary Bose-Einstein condensates, confined in a ring-shaped waveguide. During their time-evolution, the two condensate species generally remain miscible, as observed in the spatial density distributions and the autocorrelation functions. Although, the investigation is carried out for a wide range of mass-imbalance, initial demonstration is focussed on insignificant mass-imbalance of the two Rb-isotopes with suitable experimental parameters. The characteristic time scales are influenced by the trap parameters and the strengths of nonlinearities. The study also reveals the conditions under which the condensates become spatially distinguishable with clear signatures in their autocorrelation functions. A separability function further identifies favorable parameters and the fractional revival instances for greater separability. We report precise range of the ring-radius and the interaction strength for experimental realization. Additionally, the average separability variation reflects the result across a variety of condensate species. △ Less

Submitted 8 February, 2025; v1 submitted 18 September, 2023; originally announced September 2023.

Comments: 24 pages, 8 figures

Effects of equatorially-confined shear flow on MRG and Rossby waves

Authors: Mukesh Singh Raghav, Sharath Jose, Amit Apte, Rama Govindarajan

Abstract: Linear modal stability analysis of a mean zonal shear flow is carried out in the framework of rotating shallow water equations (RSWE), both under the $β$-plane approximation and in the full spherical coordinate system. Two base flows -- equatorial easterly (EE) and westerly (EW) -- with Gaussian profiles highly confined to small latitudes are analyzed. At low Froude number, mixed Rossby-gravity (M… ▽ More Linear modal stability analysis of a mean zonal shear flow is carried out in the framework of rotating shallow water equations (RSWE), both under the $β$-plane approximation and in the full spherical coordinate system. Two base flows -- equatorial easterly (EE) and westerly (EW) -- with Gaussian profiles highly confined to small latitudes are analyzed. At low Froude number, mixed Rossby-gravity (MRG) and Rossby waves are found to be particularly affected by shear, with prominent changes at higher wavenumbers. These waves become practically non-dispersive at large wavenumbers in EE. The perturbations are found to be more confined equatorially in EE than in EW with the degree of confinement being more pronounced in the $β$-plane system compared to the full spherical system. At high Froude number, the phase speeds are significantly larger in the $β$-plane system for all families of waves. Under the $β$-plane approximation, exponentially unstable modes can be excited, having negative (positive) phase speed in EE (EW). Strikingly, this flow is always neutrally stable with the full spherical system. This speaks for the importance of studying the whole spherical system even for equatorially confined shear. △ Less

Submitted 12 October, 2022; originally announced October 2022.

Journal ref: Dynamics of Atmospheres and Oceans 100 (2022) 101331

Tunneling and Revival of Anderson Localization in Bose-Einstein Condensate

Authors: Sriganapathy Raghav, Barun Halder, Pradosh Basu, Utpal Roy

Abstract: We provide an analytical model to fabricate an exponential localization of a Bose-Einstein condensate under bichromatic optical lattice. Such localization is famously known as Anderson localization. The degree of localization is investigated by the Participation Ratio to recognize the laser parameter domain for Anderson localization. The exponential nature of the localization is proved, where we a… ▽ More We provide an analytical model to fabricate an exponential localization of a Bose-Einstein condensate under bichromatic optical lattice. Such localization is famously known as Anderson localization. The degree of localization is investigated by the Participation Ratio to recognize the laser parameter domain for Anderson localization. The exponential nature of the localization is proved, where we also identify the Localization Length. The tunneling of Anderson-localized condensate with time is observed, and the revival phenomenon of Anderson localization is reported. Slowing down of Anderson localization is noticed for higher laser intensity. We also study the dynamical and structural stability of the condensate during Anderson localization, which suggests the preferred values of laser power and time instance to encounter minimal mean difference in the presence of noise. △ Less

Submitted 25 November, 2022; v1 submitted 5 August, 2022; originally announced August 2022.

Risk Assessment Techniques and Survey Method for COTS Components

Authors: Rashmi Gupta, Shalini Raghav

Abstract: The Rational Unified Process a software engineering process is gaining popularity nowadays. RUP delivers best software practices for component software Development life cycle It supports component based software development. Risk is involved in every component development phase .neglecting those risks sometimes hampers the software growth and leads to negative outcome. In Order to provide appropri… ▽ More The Rational Unified Process a software engineering process is gaining popularity nowadays. RUP delivers best software practices for component software Development life cycle It supports component based software development. Risk is involved in every component development phase .neglecting those risks sometimes hampers the software growth and leads to negative outcome. In Order to provide appropriate security and protection levels, identifying various risks is very vital. Therefore Risk identification plays a very crucial role in the component based software development This report addresses incorporation of component based software development cycle into RUP phases, assess several category of risk encountered in the component based software. It also entails a survey method to identify the risk factor and evaluating the overall severity of the component software development in terms of the risk. Formula for determining risk prevention cost and finding the risk probability is also been included. The overall goal of the paper is to provide a theoretical foundation that facilitates a good understanding of risk in relation to componentbased system development △ Less

Submitted 13 February, 2012; originally announced February 2012.

Comments: International Journal of Software Engineering & Applications (IJSEA), Vol.3, No.1, January 2012